Breadcrumb
Search and Filter
The Repository
0-3 years
Research synthesis is AI-generated, human reviewed. Updated 11/2026.
Displaying 1 - 20 of 20
The Policy Deficit In AI X Social-Emotional Learning Research
Tran Van Cuong, Liu Yihan, Nguyen Van Tuong. (08/2026). arXiv.
Nearly three-quarters of reviewed studies (72%) did not mention any substantive consequences or implications for policy governance.
Where'S The Structure? A Systematic Literature Review Of Empirical Research On Human-AI Collaboration And Hybrid Intelligence For Learning
Luis Pablo Prieto, Juan Ignacio Asensio-Pérez, María Jesús Rodríguez-Triana, Mohamed Saban, Yannis Dimitriadis. (05/2026). arXiv.
Most studies relied on basic interaction setups, primarily where AI creates or answers while humans refine (46 papers) or humans create while AI assesses (18 papers).
Pacee: Parent-Centered AI Scaffolding For Emotion Education In Early Childhood Conversations
Yu Mei, Xutong Wang, Ziyao Zhang, Yiming Fu, Shiyi Wang, Qingyang Wan, Qinghuan Lan, Chang Liu, Jie Cai, Chun Yu, Yuanchun Shi. (03/2026). arXiv.
PACEE enhanced parent-child engagement and fostered deeper emotional communication during family conversations.
Beyond Accuracy: Towards A Robust Evaluation Methodology For AI Systems For Language Education
James Edgell, Wm. Matthew Kennedy, Isaac Pattis, Ben Knight, Danielle Carvalho, Elizabeth Wonnacott. (03/2026). arXiv.
Evaluators validated pilot tasks as authentic teaching scenarios with a mean rating of 4.24 out of 5, but scored evaluation criteria significantly lower at 3.93.
Pedaco-Gen: Scaffolding Pedagogical Agency In Human-AI Collaborative Video Authoring
Injun Baek, Yearim Kim, Nojun Kwak. (02/2026). arXiv.
Videos created with PedaCo-Gen showed an average quality score increase from 3.07 to 3.86 out of 5 across multimedia learning principles.
Sketch2Feedback: A Grammar-In-The-Loop Framework For Rubric-Aligned Feedback On Student Stem Diagrams
Aayam Bansal. (02/2026). arXiv.
The end-to-end model detected errors better on free-body diagrams (micro-F1 0.471 vs. 0.263), while the grammar pipeline substantially outperformed on circuits (0.329 vs. 0.038).
Learning In Focus: Detecting Behavioral And Collaborative Engagement Using Vision Transformers
Sindhuja Penchala, Saketh Reddy Kontham, Prachi Bhattacharjee, Nima Mahmoodi, Daniel Fonseca, Sareh Karami, Mehdi Ghahremani, Andy D. Perkins, Shahram Rahimi, Noorbakhsh Amiri Golilarz. (12/2025). arXiv.
The Swin Transformer achieved the highest overall performance, reaching 97.58% accuracy, 0.9805 recall, and an F1-score of 0.9701.
Sproutbench: A Benchmark For Safe And Ethical Large Language Models For Youth
Wenpeng Xing, Lanyi Wei, Haixiao Hu, Jingyi Yu, Rongchang Li, Mohan Li, Changting Lin, Meng Han. (12/2025). arXiv.
Models showed substantial safety gaps, with large models over 30 billion parameters consistently performing well while tiny models under 500 million were 2.45 times overrepresented among bottom performers.
Kidspeak: A General Multi-Purpose LLM For Kids' Speech Recognition And Screening
Rohan Sharma, Dancheng Liu, Jingchen Sun, Shijie Zhou, Jiayu Qin, Jinjun Xiong, Changyou Chen. (12/2025). arXiv.
KidSpeak with a phonetically trained Whisper encoder achieved an average accuracy of 87.0% across four pediatric speech classification and transcription tasks.
Pacee: Supporting Children'S Personal Emotion Education Through Parent-Ai Collaboration
Yu Mei, Xutong Wang, Ziyao Zhang, Yiming Fu, Shiyi Wang, Qingyang Wan, Qinghuan Lan, Chang Liu, Jie Cai, Chun Yu, Yuanchun Shi. (11/2025). arXiv.
PACEE significantly enhanced parent-child engagement during emotional learning.
Quantum Annealing For Staff Scheduling In Educational Environments
Alessia Ciacco, Francesca Guerriero, Eneko Osaba. (10/2025). arXiv.
On the real-world school instance, the hybrid quantum annealing solver reached the optimal schedule found by the exact solver across all five runs in about 15.16 seconds.
A Meta-Analysis Of Llm Effects On Students Across Qualification, Socialisation, And Subjectification
Jiayu Huang, Ruoxin Ritter Wang, Jen-Hao Liu, Boming Xia, Yue Huang, Ruoxi Sun, Jason (Minhui) Xue, Jinan Zou. (09/2025). arXiv.
Large language models showed significant positive overall effects on academic qualification (g = 0.751), socialisation (g = 0.745), and learner subjectification (g = 0.654).
Benchmarking the Pedagogical Knowledge of Large Language Models
Maxime Lelievre, Amy Waldock, Meng Liu, Natalia Valdes Aspillaga, Alasdair Mackintosh, María José Ogando Portela, Jared Lee, Paul Atherton, Robin A. A. Ince, Oliver Garrod. (07/2025). arXiv.
Across 97 evaluated models, accuracy on general pedagogical knowledge ranged from 28% for Llama-3.2 1B to 89% for Gemini-2.5 Pro.
A Comprehensive Review on Generative AI for Education
Uday Mittal, Siva Sai, Vinay Chamola, Devika Sangwan. (09/2024). IEEE.
Generative AI can create diverse instructional materials, including texts, images, and videos, to support education.
Large Language Models for Education: A Survey
Hanyi Xu, Wensheng Gan, Zhenlian Qi, Jiayang Wu, Philip S. Yu. (05/2024). arXiv.
Large language models empower education through personalized learning support, real-time tutoring, automated assessment feedback, and teacher workload alleviation across multiple subjects.
Realizing Visual Question Answering for Education: GPT-4V as a Multimodal AI
Gyeong-Geon Lee, Xiaoming Zhai. (05/2024). arXiv.
When assessing five student hand-drawn science models, GPT-4V achieved 60% accuracy for particle structure and 80% accuracy for particle distribution compared to human coders.
AI AND PERSONALIZED LEARNING: BRIDGING THE GAP WITH MODERN EDUCATIONAL GOALS
Kristjan-Julius Laak, Jaan Aru. (04/2024). arXiv.
Current adaptive learning systems prioritize short-term test performance and domain-specific knowledge over long-term conceptual understanding, learner agency, and general competencies.
Foundation Models in Augmentative and Alternative Communication: Opportunities and Challenges
Ambra Di Paola, Serena Muraro, Roberto Marinelli, Christian Pilato. (01/2024). arXiv.
General text-to-image models like DALL-E can generate visual symbols, but their outputs are often too visually complex for most communication disability users.
Data-Driven Artificial Intelligence in Education: A Comprehensive Review
Kashif Ahmad, Waleed Iqbal, Ammar El-Hassan, Junaid Qadir, Driss Benhaddou, Moussa Ayyash, Ala Al-Fuqaha. (09/2023). IEEE.
Student grading and evaluation represents the most widely explored educational application of data-driven artificial intelligence.
Artificial Intelligence in Education: A Review
Lijia Chen, Pingping Chen, Zhijian Lin. (04/2020). IEEE.
Educational institutions have extensively adopted and used artificial intelligence, evolving from basic computers to web-based intelligent systems, embedded systems, humanoid robots, and chatbots.
